the activity of the Na⁺-K⁺-ATPase (NKA) can modulate Ca²⁺ homeostasis in vascular smooth muscle cells (VSMCs) through its influence on the electrochemical Na⁺ gradient and, thereby, cell membrane potential and the activity of the Na⁺-Ca²⁺ exchanger (NCX). Alterations in NKA activity through either inhibition with cardiac glycosides (5) or genetic manipulation in mouse models (10, 11) lead to changes in cell Ca²⁺ homeostasis and, in some cases, an enhanced loading of intracellular Ca²⁺ stores (5, 10). An explanation for the enhanced store loading is that the elevation in the intracellular Na⁺ concentration ([Na⁺]_i) slows the clearance of Ca²⁺ from the cytosol via NCX and thereby allows the sarco(endo)plasmic reticulum Ca²⁺-ATPase (SERCA) pump to sequester more Ca²⁺ into stores. Moreover, there appears to be selectivity of this process with respect to the exact isoform of the NKA that is involved in the store-loading mechanism (10). The α₁-isoform of NKA, which accounts for about 70% of the total NKA protein in mouse aortic smooth muscle (MASM) (27), has a low affinity [dissociation constant (K_d), ∼5 × 10⁻⁵ M] for cardiac glycosides, whereas the α₂-isoform, which accounts for the remainder of the NKA expression, has a relatively high affinity (4 × 10⁻⁸ M) for cardiac glycosides (19). Since concentrations of cardiac glycosides that only affect the activity of the α₂-isoform of NKA (50 μM) enhance store loading in rat mesenteric myocytes (2), and the hyperloading of stores is also observed in astrocytes from NKA α₂-isoform gene-ablated, homozygous null knockout (α₂-KO) mice (10), it appears possible that this specific isoform of the NKA is coupled to NCX and to the loading of Ca²⁺ stores.

A complicating factor regarding the observed effects of ouabain on the activity of the α₂-NKA relates to the specificity of ouabain for inhibiting NKA activity alone. It has been known for several years that low concentrations of ouabain alter the proliferative state of vascular smooth muscle (3). More recently, this response has been delineated to an activation of Src kinase signaling, which is dependent on molecular interactions between Src and the NKA (29). Treatment with ouabain is purported to activate Src, subsequently leading to the generation of inositol 1,4,5-trisphosphate [Ins(1,4,5)P₃] and the activated release of Ca²⁺ from intracellular stores (29, 33). To address the complexity of the potential effects on Ca²⁺ signaling and store function in response to the inhibition of the α₂-NKA, Edwards and Pallone (9) recently presented a mathematical model, which takes into account these competing effects, and evaluated possible outcomes of low-dose ouabain on Ca²⁺ homeostasis. The conclusions drawn from this model indicate that the inhibition of α₂-NKA favors store loading, whereas the activation of Src and, thereby, increases in Ins(1,4,5)P₃ production favor store depletion. Based on these considerations, the primary effect of ouabain in modulating vascular smooth muscle Ca²⁺ homeostasis is likely via the inhibition of the NKA and subsequent inhibition of the Ca²⁺ efflux via the NCX. Interestingly, we recently reported that force in response to receptor-mediated stimulation but not to KCl was more sensitive in aorta from the α₂-KO than from the wild-type (WT) (27) mouse, suggesting that activation via store release is enhanced in the α₂-KO mouse. Importantly, when total NKA was reduced to a similar extent in the α₁-isoform heterozygous (Het) null mouse, no changes in aortic contractility were observed. These observations are consistent with the hypothesis of hyperloaded stores due to a specific lack of the α₂-isoform.

In the present work, we investigated the effects of gene targeting of the α₂-NKA isoform on vascular smooth muscle Ca²⁺ homeostasis by analyzing Ca²⁺-handling processes in isolated MASM cells. Specifically, we tested the hypothesis that Ca²⁺ stores are increased in MASM cells from α₂-KO mice. Since the α₂-KO mouse dies shortly after birth (11), we developed protocols to utilize cells from aorta of day 18 embryonic mice for these studies. Our results indicate that Ca²⁺ clearance is altered in the α₂-KO smooth muscle cells, whereas store Ca²⁺ loading is unchanged. Increased PMCA activity and blunted NCX activity appear to underlie the differences in Ca²⁺ clearance between WT and α₂-KO smooth muscle cells.

MATERIALS AND METHODS

Mice.

All breeding and surgical procedures were performed according to University of Arizona Institutional Animal Care and Use Committee-approved protocols. Heterozygous NKA α₂^+/− (α₂-Het) mice on the C57/BL6 background were obtained from Dr. Jerry Lingrel (University of Cincinnati). Mice heterozygous for the transgene were mated, as previously reported Shelly et al. (27). Litters were collected for cell isolation at embryonic day 18, since homozygote α₂-NKA (α₂-KO) mice die minutes after birth.

Aorta and cell isolation.

Day 18 mouse embryos were obtained from females heterozygous for the α₂-isoform that were mated with male mice of the same genotype. These mice are on a C57/Black Swiss background. The dam was euthanized with CO₂, and the placenta was placed into ice-cold saline. Each embryo was removed and euthanized by decapitation, followed by the removal of the entire thoracic aorta (exposed from aortic arch to diaphragm). After the aorta was isolated, a tail clip was obtained and stored for subsequent genomic analysis. The aorta was dissected free from all associated fatty tissue and removed by cutting the intercostal arteries. The blood was immediately flushed from the isolated vessel with PBS + 0.1% BSA, and then the aorta was placed in 1 ml of PBS containing collagenase B (6 mg/ml) and BSA (6 mg/ml) and incubated at 37°C for 40 min. Cells were dispersed from the digested aorta by trituration, diluted in PBS, and centrifuged at ∼1,000 g for 3 min. The cell pellet was suspended in 250 μl of DMEM supplemented with 5% fetal bovine serum and 1% penicillin-streptomycin. The 250 μl of cell suspension from one specific aorta were plated as 50-μl aliquots onto the center of six glass coverslips housed in individual wells of a six-well plate. The six-well plate was then placed into a 37°C incubator equilibrated with 95% air–5% CO₂. After ∼16 h, 5 ml of fresh DMEM were added to each well. The cells were used for experiments between 24 and 48 h. All functional (Ca²⁺) experiments were performed before genotyping (i.e., blind). The PBS contained (in mM/l) 2.7 KCl, 1.5 KH₂PO₄, 138 NaCl, and 8 Na₂HPO₄ at pH 7.2.

Cytosolic Ca²⁺ measurements.

For cytosolic Ca²⁺ ([Ca²⁺]_i) measurements, coverslips containing MASM cells were washed with HBSS containing (in mM) 5 KCl, 0.3 KH₂PO₄, 138 NaCl, 0.2 NaHCO₃, 0.3 Na₂HPO₄, 20 HEPES, 1.3 CaCl_2, 0.4 MgSO₄, and 5.6 glucose (pH 7.3) at 37°C for 30 min and then loaded with 2.5 μM of the acetoxymethylester form of fura-PE3 (Teflabs, Austin, TX) with 0.0025% pluronic acid in HBSS for 20 min at 37°C. The cells were rinsed in HBSS for 20 min at 37°C to allow for the hydrolysis of the fura-2 AM. The coverslip with the dye-loaded cells was then placed in a chamber held at 37°C while mounted on the stage of an inverted Olympus IX-70 microscope equipped with a 40 × 1.4 numerical aperature (NA) ultrafluor objective and a 150-W Xe lamp as the excitation source. Fura-PE3 loaded within the cells was alternately excited at 340 and 380 nm using a filter wheel. The emitted light was filtered at 510 nm (10 nm band pass) before focusing the cell image onto a charge-coupled device (CCD) camera (Photometrics CH-250). The collection time for an image pair was 2 s. To evaluate absolute Ca²⁺ concentrations and rates of Ca²⁺ flux, the fura affinity for Ca²⁺ was evaluated from an in situ calibration procedure (16). An averaged K_d of fura-PE3 for Ca²⁺ of 0.48 μM with a maximum fura-2 ratio (R_max) of 4 and a minimum fura-2 ratio (R_min) of 0.29 was obtained from these measurements. Measured ratios of the 340 and 380 intensities were used to estimate [Ca²⁺]_i by the equation:

To correct for differences in absolute ratio values observed in individual cells and allow for the analysis of relative changes in Ca²⁺ in response to an effector, a standard normalization scheme was utilized. To initiate an experiment, cells were perfused in a nominally Ca²⁺-free media (free Ca²⁺, ∼5 μM as measured by Mag-fura-2) for 1.5 min before the release of Ca²⁺ from stores. These time course data are presented in Fig. 1. To compare changes in [Ca²⁺]_i between experiments, peak values acquired after a perturbation from an individual cell were normalized to the last measurement in the nominally Ca²⁺-free media before perturbation.

Immunocytochemistry.

The cells on coverslips, 24–48 h after isolation, were fixed with 3% paraformaldehyde, rinsed with 25 mM glycine, and permeabilized with 0.1% Triton X-100 (15). The cells were incubated with the primary antibodies against the α₁- or α₂-isoforms of NKA (both rabbit polyclonal antibodies) (22) and antibodies (both monoclonal) raised against the rat NCX (Swant, Bellinzona, Switzerland) and PMCA (Research Diagnostics) pumps for 2 h at 25°C. Cells were washed three times (10 min each) in PBS to remove unbound primary antibody and then incubated with a secondary anti-rabbit IgG labeled with Texas red (Invitrogen) and anti-mouse IgG labeled with Alexa 530 (Invitrogen) for 45 min at 25°C. Secondary antibody-only controls were run for all combinations. Two controls were also used to determine the specificity of the α₂-NKA antibody (HERED). First, the antibody was incubated with the HERED antigen peptide [×10 peptide (wt/wt)] for 60 min before incubation with cells isolated from WT mice. In addition, the treatment of cells isolated from α₂-KO mice served as controls for antibody specificity.

For samples where the distributions of both α₁- and α₂-isoforms were evaluated, an intermediate blocking step was required (12). Briefly, cells were labeled with one primary and then a fluorophore-conjugated secondary anti-rabbit antibody, as described above. The sample was next incubated with an unlabeled anti-rabbit antibody at an approximately ×10 concentration (0.1 mg/ml) overnight (at least 10 h) before labeling with the second set of antibodies. Cross-reactivity was tested by leaving out the second primary antibody but adding the second fluorescent anti-rabbit antibody (14). If any significant labeling was observed with the second fluorophore, the preparation was not acceptable. All coverslips were mounted onto glass slides using a 50% glycerol-saline solution containing the antibleach agent paraphenylendiamine (0.1%).

For standard wide-field imaging, the slides were mounted onto the stage of an Olympus IX-70 microscope equipped with a 60 × 1.4 NA objective. Illumination was provided by a 100-W Hg lamp, and images were acquired using a liquid-cooled CCD camera (Roper Scientific) equipped with a Kodak CCD array (KAF1401E). Three-dimensional image acquisition and deconvolution were performed using a DeltaVision restoration microscopy system (Applied Precision, Issauah, WA). Image deconvolution was carried out using an iterative approach based on regularization (7). All other image analyses were performed on Silicon Graphics Indy2 workstations using customized software. The analysis of colocalization between images collected for α₂-NKA and NCX distributions was performed as described previously (23).

Statistical analyses.

The experimental data are expressed as means ± SE and were analyzed using an unpaired Student's t-test. P values of <0.05 were taken to indicate statistical significance.

Materials.

Fura-2 PE3-AM and ionomycin were purchased from Teflabs (Austin, TX). All other chemicals were of reagent grade and were purchased from Sigma (St. Louis, MO).

RESULTS

To functionally evaluate the transport mechanisms responsible for maintaining Ca²⁺ homeostasis and to compare responses between WT and α₂-KO mice, we employed a strategy to perturb [Ca²⁺]_i and then estimate the contributions of the various pathways of Ca²⁺ clearance. Ca²⁺ stores were unloaded using cyclopiazonic acid (CPA) to pharmacologically inhibit the SERCA pumps and thereby permit Ca²⁺ efflux through leak pathways to empty all SERCA-containing compartments. Ca²⁺ influx was limited by the incubation of cells in a nominally Ca²⁺-free medium.

The standard protocol used to evaluate Ca²⁺ handling is shown in Fig. 1. Cells were perfused in a nominally Ca²⁺-free media for 1.5 min before the addition of CPA to limit the contribution of Ca²⁺ influx after store unloading with CPA. When Ca²⁺ was removed from the incubation media, a rapid decline in [Ca²⁺]_i was observed. The averaged [Ca²⁺]_i in both the presence and absence of media Ca²⁺ is shown in Fig. 2. [Ca²⁺]_i was ∼65 nM in the presence of 1.6 mM extracellular Ca²⁺ and 31 nM after Ca²⁺ was removed from the media. There were no significant differences in the values between cells isolated from WT, α₂-Het, or α₂-KO mice. If Ca²⁺ was added back to the media 1.5 min after its removal without any other perturbation, [Ca²⁺]_i rapidly (within 15 s) recovered to its original baseline. If CPA was added 1.5 min after the Ca²⁺ removal to initiate the unloading of Ca²⁺ stores, a slow-developing and relatively small elevation in [Ca²⁺]_i was observed (Fig. 1). The subsequent elevation of medium Ca²⁺ to 1.6 mM in the presence of CPA elicited a rise in [Ca²⁺]_i that exceeded the original resting [Ca²⁺]_i purportedly due to the opening of store-operated channels and the subsequent capacitative calcium entry (CCE), as previously described (4, 32). The magnitude of the rise in [Ca²⁺]_i after either CPA or the elevation of medium Ca²⁺ is dependent on the combination of store release and clearance of Ca²⁺ from the cytosol.

Mechanisms for Ca²⁺ clearance.

As seen in Fig. 3, the peak increase in [Ca²⁺]_i elicited by CPA is less in cells from α₂-KO relative to WT mice. Since changes in cytosolic Ca²⁺ clearance could attenuate the magnitude and rate of rise of the apparent Ca²⁺ release, experiments were performed to evaluate whether differences in the clearance by several different processes could account for the lower apparent Ca²⁺ release observed in the α₂-KO cells. Blocking the activity of the NCX, by the isosmotic substitution of extracellular Na⁺ with N-methyl-glucosamine, enhanced the CPA-induced Ca²⁺ transient observed in cells from WT and Het mice but did not significantly increase the apparent Ca²⁺ release in cells from α₂-KO mice (Figs. 1 and 3). Thus, in cells from WT mice, NCX activity greatly attenuates the magnitude of the peak of the Ca²⁺ transient caused by store unloading with CPA. This clearly is not the case for cells from α₂-KO animals.

The influence of the plasma membrane Ca²⁺-ATPase (PMCA) was assessed by incubation with vanadate, a nonspecific P-type ion ATPase inhibitor. As shown in Fig. 4, vanadate by itself greatly enhanced the Ca²⁺ transient in response to CPA in cells from α₂-KO mice but did not enhance the effect in WT cells. Thus these data argue that Ca²⁺ released from intracellular stores is preferentially removed from the cytosol via PMCA in cells from α₂-KO mice but by NCX in cells from WT mice. When both NCX and PMCA were inhibited, there was no significant difference in the Ca²⁺ release elicited by CPA (Fig. 4).

The other Ca²⁺ transport mechanism that could influence the magnitude of the CPA-induced Ca²⁺ transient is the mitochondrial Ca²⁺ uniporter (21). However, this transporter may not play a significant role in attenuating the CPA response because its affinity for Ca²⁺ is relatively low [Michaelis-Menten constant (K_m), >5 μM]. The role of mitochondria in modulating the magnitude of the CPA response was evaluated in WT cells by adding a protonophore FCCP immediately before CPA. FCCP rapidly dissipates the mitochondrial membrane potential and thereby blocks the mitochondrial Ca²⁺ uptake mechanism. FCCP did not enhance the observed Ca²⁺ release in response to CPA above that seen with Na⁺-free media alone, indicating that mitochondria do not modulate Ca²⁺ transients (Fig. 5), at least in the case of a CPA-induced response, which does not reach 100 nM (Fig. 1). Similar findings were made if cells were incubated with 10 μM ruthenium red, a blocker of the Ca²⁺ uniport mechanism, indicating that mitochondrial sequestration does not significantly alter the magnitude or shape of the CPA-induced release of store Ca²⁺.

Based on our findings shown in Figs. 2–5, the total CPA-releasable store Ca²⁺ can be assessed by inhibiting the Ca²⁺ clearance pathways such as NCX with Na⁺-free solution and PMCA with vanadate (1 μM). Therefore, we performed a series of experiments where Ca²⁺ clearance was inhibited to evaluate the magnitude of store loading in cells from WT relative to α₂-KO mice. Although the observed CPA-releasable Ca²⁺ without the inhibition of Ca²⁺ clearance was considerably lower in the α₂-KO than WT (Fig. 2), when measured with Ca²⁺ clearance inhibited, the total releasable Ca²⁺ within the stores was similar (Fig. 6).

Differences in the permeability of the stores to Ca²⁺ (P_SR) could influence the interpretation of the peak Ca²⁺ responses used to estimate the relative contribution of the various clearance components. To estimate the leak permeability, we measured the initial rate of [Ca²⁺]_i increase elicited by CPA in the presence of inhibitors. These measured initial rates, in WT, 44.5 ± 4.3 nmol/s (n = 48), versus α₂-KO, 39 ± 1.9 nmol/s (n = 22), were not different. Since the Ca²⁺ leak rate is proportional to store Ca²⁺ load ([Ca²⁺]_SR × P_SR) and the measured total [Ca²⁺]_SR was equal in WT and α₂-KO cells, the Ca²⁺ leak permeability must also be equal. Our present data and previous studies from Blaustein and colleagues indicate that NCX is functionally coupled to the α₂-isoform of NKA. Moreover, Blaustein and colleagues (2, 10) have reported that the inhibition of the α₂-isoform by low concentrations of ouabain or gene ablation leads to the enhanced loading of Ca²⁺ stores in rat mesenteric myocytes and mouse astrocytes, respectively. Although we directly inhibit NCX by the removal of extracellular Na⁺, this was done acutely (1.5 min before store release) to have minimal effect on store loading. Therefore, to test whether selective α₂-NKA inhibition for longer periods alters store loading, we incubated cells with 0.5 μM ouabain for 20 min and then evaluated store load by treating cells with CPA to release Ca²⁺ in the presence of Ca²⁺ clearance inhibitors. As shown in Fig. 7, the treatment of cells from WT mice with ouabain for 20 min before initiating the standard protocol enhanced the CPA-induced Ca²⁺ transient relative to that observed in cells incubated only with inhibitors of Ca²⁺ efflux. However, when cells from α₂-KO mice were similarly treated with ouabain and then analyzed for store load in the presence of efflux inhibitors, the amount of Ca²⁺ released from the stores by CPA was not different than that seen in cells incubated in the presence of efflux inhibitors alone.

Calcium entry.

Although the amount of Ca²⁺ released from stores can be appreciable, the subsequent influx of Ca²⁺ initiated by store release is of much greater magnitude (6, 8). As shown in the standard protocol (Fig. 1), several minutes after the CPA-induced response, Ca²⁺ was added to evaluate the magnitude of the increase in [Ca²⁺]_i, as an estimate of influx via Ca²⁺ channels. In the absence of Ca²⁺ clearance inhibitors, the magnitude of the [Ca²⁺]_i increase also appeared to be significantly although not substantially depressed in cells from the α₂-KO mice (Fig. 8). The acute inhibition of NCX substantially enhanced the Ca²⁺ influx-related increase in [Ca²⁺]_i in cells from both WT and α₂-KO mice. Since acute Na⁺-free conditions enhanced the increase in [Ca²⁺]_i in cells from α₂-KO mice, NCX is in fact active in these cells. The inhibition of both NCX and PMCA further enhanced the peak [Ca²⁺]_i increase in both WT and α₂-KO cells, indicating that PMCA activity is also important for depressing the CCE-linked Ca²⁺ transient. However, the magnitude of the Ca²⁺ channel-dependent influx (with clearance inhibited) was significantly larger in cells isolated from α₂-KO mice (Figs. 8 and 9).

To evaluate the ability of mitochondria to alter the Ca²⁺ influx-dependent Ca²⁺ transient, the mitochondrial uniporter was blocked with ruthenium red (added 20 min before the initiation of the Ca²⁺ measurement protocol). With the Ca²⁺ efflux pathways also inhibited, ruthenium red further enhanced the peak of [Ca²⁺]_i reached after the readmission of Ca²⁺ (Fig. 9), suggesting that mitochondrial Ca²⁺ clearance is significant during large increases in [Ca²⁺]_i. The magnitude of the increase in [Ca²⁺]_i remained greater in cells from α₂-KO relative to WT mice when all potential clearance pathways were inhibited. SKF-96365 (10 μM) (31) was used to evaluate the dependence of the Ca²⁺-induced increase in [Ca²⁺]_i on the opening of nonselective cation channels such as store-operated channels. SKF-96365 blocked most of the increase in [Ca²⁺]_i following the addition of Ca²⁺ to the media, supporting a role for a nonselective cation channel in carrying the channel-dependent Ca²⁺ entry (Fig. 9). 2-Aminoethoxydiphenyl borate had a similar effect in inhibiting most of the Ca²⁺ entry, also supporting the role for store-operated channels (data not shown).

Subcellular distribution of α-isoforms of NKA.

The distribution of NKA α-isoforms, NCX, and Ca²⁺-ATPases was evaluated using immunocytochemistry and three-dimensional imaging techniques. As described previously in astrocytes from embryonic WT mice (13) and aortic myocytes from adult mice (27), the α₂-isoform has a restricted distribution, whereas the α₁-isoform is diffusely (uniformly) distributed over the entire cell surface of myocytes isolated from WT mice (Figs. 10 and 11 ). Nonspecific labeling was evaluated using the α₂-isoform-specific antibody on cells from α₂-KO mice or after prior incubation with a ×10 concentration of the HERED antigen peptide (22). With both approaches, no significant labeling above that seen for secondary Texas red-labeled anti-rabbit alone (background) was observed (not shown). Since our Ca²⁺ measurements suggest a functional interaction between the α₂-isoform of the NKA and NCX but not PMCA, we evaluated the distributions of these proteins.

Like the α₂-isoform, NCX distribution was highly restricted, whereas PMCA distribution resembles the α₁-NKA in that it is more uniformly distributed (Fig. 12). In cells labeled for both the α₂-isoform and NCX, a similar distribution of the two proteins was clearly observed (Fig. 13). An analysis of colocalization between images was performed to quantify the level of coincidence between the markers for the two transporters (23). Approximately 50% (± 4 SE) of α₂-NKA-positive pixels was found to be coincident with pixels identified as containing NCX. Conversely, only about 37% (± 5 SE) of NCX-positive pixels were found to be coincident with pixels identified as containing α₂-NKA (n = 3 comparisons), suggesting that more α₂-NKA is colocalized with NCX than visa versa. The extent of colocalization between these two proteins that could be attributed to chance is far lower, in the range of 1–3%, depending on estimates of total cell volume occupied by the individual markers. It should also be noted that maximum colocalization is not expected to be near 100% even if the two probes are expressed in the same compartment, due to a variety of image acquisition and analysis limitations, with a maximum anticipated coincidence of ∼70% (24). Therefore, much if not all of the membrane-localized α₂-NKA appears to be colocalized with NCX.

DISCUSSION

Several molecular components required for Ca²⁺ homeostasis and Ca²⁺ store function are proposed to be organized at specific membrane domains adjacent to Ca²⁺ stores in smooth muscle (6, 10) and potentially other cell types (6, 10). Moreover, a functional coupling between the α₂-isoform of NKA and NCX is believed to be important for modulating the loading state of Ca²⁺ stores (6, 10). Data presented here are consistent with most aspects of this model. A proposed consequence of losing this molecular coupling, and therefore loss of Ca²⁺ extrusion via the NCX, is that both intracellular Ca²⁺ concentration and Ca²⁺ store load will increase, as was observed in astrocytes isolated from the α₂-KO relative to WT mouse line (10). However, neither in the presence or absence of 1.6 mM extracellular Ca²⁺ did we observed any differences in the basal [Ca²⁺]_i in VSMCs isolated from α₂-KO, α₂-Het, and WT mice (Fig. 2). These differences between cell types isolated from the same mouse line are potentially due to differences in compensatory mechanisms or unique processes by which smooth muscle cells organize their pathways for Ca²⁺ homeostasis. Moreover, the previous studies did not inhibit Ca²⁺ clearance during their analysis of store load, thereby complicating the interpretation of the magnitude of Ca²⁺ released from stores. We designed a series of experiments to evaluate potential compensatory mechanisms for Ca²⁺ regulation in the MASM cells.

A difficulty with using isolated cells in culture is that some cells within the population lose sensitivity to natural agonists, making their response to receptor-activated processes heterogeneous, thereby complicating data analysis when comparing responses of cells within two populations. To minimize the effects of cell heterogeneity, a standardized protocol was designed to evaluate the regulation of [Ca²⁺]_i (Fig. 1). To initiate an experiment, cells were perfused in a nominally Ca²⁺-free media (free Ca²⁺, ∼5 μM as measured by Mag-fura-2) for 1.5 min before the addition of CPA. The use of CPA to release stores, rather than a receptor-mediated agonist, circumvents the differences between individual cells in receptor expression or second messenger production, which may cause differences in receptor-mediated activation. The removal of media Ca²⁺ limits Ca²⁺ influx after store unloading. Clearance of the released Ca²⁺ must also be inhibited if the measured change in [Ca²⁺]_i is to be used as an estimate of store content. By selective inhibition of the known Ca²⁺ clearance pathways (8), the influence of each mechanism on Ca²⁺ homeostasis can also be evaluated (26). Our approach was to add the inhibitors, in the Ca²⁺-free buffer 1.5 min before the release of Ca²⁺ stores (with CPA), to limit their effects on the level of stored Ca²⁺. The activity of the NCX was inhibited by the isosmotic substitution of extracellular Na⁺ with N-methyl-glucosamine. The influence of the PMCA was assessed by the incubation with vanadate (30), a nonspecific P-type ion ATPase inhibitor. Vanadate can also inhibit SERCA and H⁺ pumps; however, with our protocol, SERCA was already inhibited by CPA and no effect of vanadate on cell pH was observed over the experimental time course (data not shown). Another consideration in using vanadate is its ability to permeate the cell (25); however, treatment with vanadate clearly altered Ca²⁺ clearance, indicating at least partial effects on PMCA. Thus, with the caveat that no inhibitor is perfectly specific, our evidence suggests that the effects of vanadate reported here reflect the inhibition of PMCA.

A remaining mechanism that could influence the magnitude of the CPA-induced Ca²⁺ transient is mitochondrial Ca²⁺ uptake, which can be evaluated by adding a protonophore FCCP or using ruthenium red, a blocker of the mitochondrial Ca²⁺ uniporter. Our data suggest that mitochondria do not play a significant role in clearing Ca²⁺ from the cytosol after its release from stores under the conditions used in our experiments. Therefore, when NCX and PMCA efflux pathways are blocked, the magnitude of the CPA-induced increase in [Ca²⁺]_i is a reasonable measure of the total stored (releasable) Ca²⁺. Based on this logic, there was no apparent difference in the level of stored Ca²⁺ in cells isolated from α₂-KO relative to WT mice, as seen in Fig. 6. The rate of Ca²⁺ release from stores under these conditions was also similar in α₂-KO to WT mice. Thus our findings indicate that the store leak permeability was also similar, validating the use of peak Ca²⁺ to assess the individual Ca²⁺ clearance components. Based on this logic, the relations between the expression of the NKA and Ca²⁺ homeostasis were investigated in VSMCs isolated from WT and α₂-KO mice.

The estimation of Ca²⁺ release from stores by CPA without inhibiting clearance suggested that Ca²⁺ store levels were somewhat lower in cells from α₂-KO mice. Inhibition of NCX by itself significantly increased the peak Ca²⁺ observed after the store release by CPA in MASM cells from WT mice, whereas further inhibition of PMCA (or PMCA by itself) had little effect (Figs. 4 and 5). We interpret these data to indicate that NCX limits the observed CPA-induced Ca²⁺ transient because it has preferential access to store-released Ca²⁺, as previously suggested by van Breemen and colleagues (1, 18), who proposed a unique and distinct coupling between store Ca²⁺ and NCX. A structural basis for this coupling has also been suggested in smooth muscle (17) and astrocytes (12).

Although PMCA does not have access to the store-released Ca²⁺ in cells from WT mice, PMCA activity is substantial, since larger global elevations in [Ca²⁺]_i, as observed during the subsequent initiation of Ca²⁺ influx, are substantially enhanced by the inhibition of Ca²⁺ clearance via PMCA (Fig. 8). Of interest here is the comparison of the known affinities of the two transport processes for Ca²⁺. PMCA has a K_m of <0.5 μM (20), whereas the affinities of NCX and the mitochondrial uniporter for Ca²⁺ are near or above 1 μM (20). Therefore, the ability of NCX to act as the primary mechanism for the removal of store-released Ca²⁺, when cytosolic Ca²⁺ concentrations do not rise above 250 nM (see Fig. 1 for an example), suggests that either NCX resides in a region of the cell where the Ca²⁺ concentration is much higher than that observed with the cytosolic reporter probe (Figs. 1 and 2) or that its affinity for Ca²⁺ is enhanced in situ. PMCA, on the other hand, plays an important role in reducing the Ca²⁺ load during higher transitions in cell Ca²⁺, well within its predicted range of response. Similarly, mitochondrial Ca²⁺ sequestration becomes significant during large elevations in cell Ca²⁺, as occur after the elevation of medium [Ca²⁺] (Fig. 9). Thus, other than during small Ca²⁺ transients observed with store release, the predicted hierarchy of Ca²⁺ clearance mechanisms, based on the known affinity for Ca²⁺, is observed.

In contrast to the WT cells, cells isolated from α₂-KO mice exhibited a much different pattern of efflux activity. In these cells, the inhibition of NCX had little effect on the removal of Ca²⁺ released from stores, suggesting that NCX no longer has preferential access to this compartment. As shown in Fig. 3, incubation in Na⁺-free medium did not enhance the CPA response, whereas the inhibition of PMCA-mediated efflux significantly increased the peak Ca²⁺ response (Fig. 4). Importantly, NCX activity was not lost in the α₂-KO cells in that its inhibition enhanced the peak Ca²⁺ reached after initiating Ca²⁺ influx. Presumably, this is due to the remaining α₁-isoform maintaining global cytosolic Na⁺ low and attributable to the higher levels of [Ca²⁺]_i attained under these conditions (see below). The lack of any effect of low concentrations of ouabain on Ca²⁺ influx (Fig. 8) is also consistent with the hypothesis that the α₁-isoform is the primary driving force for maintaining the NCX extrusion of [Ca²⁺]_i when cytosolic [Ca²⁺]_i is high.

Our observation of equivalent store Ca²⁺ content and resting [Ca²⁺]_i differs from that reported for astrocytes isolated from the α₂-KO mice in which both parameters were elevated relative to the WT mice (10). However, we also found that the specific inhibition of the α₂-isoform of the NKA by treatment with 0.5 μM ouabain for 20 min before adding CPA increased store load in cells from WT mice (Fig. 7). This observation is consistent with these previous studies, suggesting that a blockade of the α₂-isoform, and thereby depression of the localized Ca²⁺ efflux via NCX, does lead to enhanced store Ca²⁺ content. This effect was not seen in cells isolated from the α₂-KO mice. Thus the absence of the α₂-isoform, an isoform that accounts for <30% of the total expressed NKA (27, 28), leads to the loss of preferential access of NCX to store-released Ca²⁺. Our results indicate that this isoform also at least partly controls the loading of Ca²⁺ stores, based on the finding with 0.5 μM ouabain inhibition of the α₂-isoform in WT cells (Fig. 7). Insight into the potential outcomes of the loss of α₂-NKA chronically (KO), as opposed to the inhibition using low doses of (<1.0 μM) ouabain, may be gained from the model presented by Edwards and Pallone (9). In this model, both the direct effects of ouabain on Ca²⁺ store load caused by inhibition of NKA activity and subsequent alteration in NCX and the indirect effects via the activation of Src signaling and Ins(1,4,5)P₃ generation were evaluated. Based on our results and previous results by others (27, 28), indicating enhanced store Ca²⁺ load after treatment with ouabain, it appears that the primary mechanism of action for ouabain in the MASM is the direct effect on NKA activity. The model also predicts that [Ca²⁺]_i will reach a new plateau after ouabain treatment such that the resting [Ca²⁺]_i will be elevated. However, in the presence of a chronic loss of the α₂-isoform, we find that resting [Ca²⁺]_i is not different than that observed in cells from WT mice and that store load is normalized. Overall, these data suggest that the ablation of the NKA α₂-isoform leads to a functional diminution of NCX activity at low [Ca²⁺]_i, which elicits a functional activation of PMCA (Fig. 4), allowing cells from the α₂-KO mice to maintain an apparently normal resting [Ca²⁺]_i (Fig. 2) and normal store Ca²⁺ content (Fig. 6).

Our previous studies showed that the sensitivity for the activation of contractility by receptor-mediated agonists, which act primarily by releasing store Ca²⁺, was enhanced in aortas from the α₂-KO mice (27). Consequently, we predicted that Ca²⁺ store load would be enhanced in cells isolated from α₂-KO mice, such that more Ca²⁺ would be released at lower agonist concentrations. The equal magnitude of store-releasable Ca²⁺ in α₂-KO and WT mice (Fig. 6) and, in particular, the lower apparent increase in [Ca²⁺]_i with CPA treatment in the α₂-KO aorta argues against this hypothesis.

The direct comparison of measurements of [Ca²⁺]_i in cultured VSMCs to contractility studies in the aorta are not straightforward. It is also important to note that the experiments presented here were performed in the absence of media Ca²⁺, so that Ca²⁺ influx would not be a variable. Therefore, a potential contributing factor in the enhanced sensitivity for contractile activation could be enhanced Ca²⁺ influx and, in particular, CCE. No significant difference in the peak CCE was observed between cells from α₂-KO relative to WT mice (Fig. 8), but when Ca²⁺ clearance was inhibited, a clear elevation in Ca²⁺ influx was observed in cells from α₂-KO mice (Figs. 8 and 9). One interpretation of our overall findings is that the elevated Ca²⁺ flux rate via CCE, without a significant difference in the peak [Ca²⁺]_i attained, is more effective in terms of activating contractility in the α₂-KO aorta.

As indicated, a structural basis for a role in the coupling of the NKA α₂-isoform NCX and Ca²⁺ store signaling with agonist activation has been proposed (12, 17). This organization is consistent with our immunochemical analysis showing a remarkable coincidence in the distribution between the α₂-NKA and NCX, which is also consistent with a high degree of colocalization between these proteins observed in smooth muscle from toad stomach (17). Moreover, our current findings place the regulation of store-operated Ca²⁺ entry into this compartment in MASM as well. Conversely, both the α₁-NKA and PMCA demonstrate a much more diffuse distribution, and there is little correlation between α₁-NKA and NCX distributions. Based on these considerations, we hypothesize that an enhanced Ca²⁺ influx into the region associated with the localized α₂-isoform may be more readily accessed by the contractile apparatus. Therefore, increased flux into this compartment in the α₂-KO aortic cells due to enhanced CCE and impaired Ca²⁺ clearance may underlie the enhanced sensitivity for contractile activation in the α₂-KO aorta (27).

GRANTS

This work was supported by grants from the Canadian Institutes of Health Research and the Heart and Stroke Foundation of British Columbia and the Yukon (to E. D. W. Moore) and by National Heart, Lung, and Blood Institute Grant HL-66044 (to R. J. Paul and R. M. Lynch).

FOOTNOTES

• The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.